Page 6 - 3GPP_Highlights_Issue_5_WEB

TECHNICAL HIGHLIGHTS

AI/ML FOR NR AIR INTERFACE

By Juan Montojo, Rapporteur for the RAN1-led study
on AI/ML for NR Air Interface, Qualcomm Inc.

Before Rel-18, Artificial Intelligence (AI) and In Rel-17 a RAN3-led study on further enhanced data collection
Machine Learning (ML) related projects investigated the high-level principles of RAN intelligence
enabled by AI. This project laid out the functional framework
in 3GPP focused on enabling network for RAN intelligence and the benefits of AI enabled NG-RAN
automation or data collection for various examining various use cases. The Technical Report (TR) of
this study can be found in 37.817 and constitutes an excellent
network functions. reference for the findings of the project. This study led to the
approval of a Rel-18 normative project on AI/ML for NG-RAN
The Network Data Analytics Function (NWDAF) was introduced focusing on enhancements to data collection and signaling to
in Rel-15 providing network slice analysis capabilities. It was support AI/ML based Network Energy Savings, Load Balancing
later expanded to providing data collection and exposure in 5G and Mobility Optimizations.
core in Rel-16, and to enable UE application data collection
in Rel-17. The Rel-18 RAN1-led study on AI/ML for NR Air Interface, as
the central subject of this article, explores the benefits of
Similarly, projects on Self Organizing Network (SON) and augmenting the air interface with features enabling improved
Minimization of Drive Tests (MDT) have been defining data support of AI/ML based algorithms for enhanced performance
collection procedures for various NR features over releases and/or reduced complexity or overhead.
starting from Rel-16. How the network would use that
collected data has always been left to implementation.

The project description has identified three promising areas which will be used as a pilot to deepen the
understanding of the solution space and corresponding performance evaluation comparisons with pertinent
non-AI/ML based implementations and across companies:

• Channel State Information (CSI) and including descriptions on training, inference, testing, and
For CSI enhancements, frequency domain compression has verification of the models. All those concepts will have to be
already been agreed, with other enhancements, e.g., time- investigated in light of their exposure to 3GPP specifications.
domain prediction, being still considered.

• Beam Management (BM) The ultimate objective of this study is
Spatial and temporal prediction seem to be promising areas
of focus. the characterization of the specification
• Positioning impact that will enable the deployment
Direct AI/ML positioning (e.g., fingerprinting) and AI/ML
assisted positioning (e.g., the output of the AI/ML model and inter-operation of these AI/ML
inference is a new measurement and/or an enhancement based techniques
of an existing measurement) are the most popular areas for
further investigation.

The AI/ML model is assumed to be running at one of the two Performance evaluations and comparisons with a meaningful
sides of the communication link, i.e., gNB or UE, for most of the non-AI/ML baseline are an integral part of the project to measure
use cases. However, the CSI use case will explore the possibility of the true potential of the AI/ML techniques. Clearly, there will
having two-sided AI/ML model with a tight interplay between the be various Key Performance Indicators (KPIs) identified for the
UE and gNB. Whether and how that interaction will be enabled different use cases. In turn, AI/ML based techniques will be
by the 3GPP is subject of discussion. identified in terms of performance and associated complexity.
This project will also identify the relevant AI/ML notation and Complexity, in addition to computational requirements, will relate
nomenclature which will be necessary for describing AI/ML to power consumption and memory utilization.
models and their life cycle in conjunction with various levels of The ultimate objective of this study is the characterization of the
collaboration between the network and the user equipment, specification impact that will enable the deployment and inter-
operation of these AI/ML based techniques.

|
06 3GP P Highlights n e w slet t er